PROJECT SUMMARY: The evolution of human cancer is a complex process driven by multiple molecular and cellular events. Cancer cells often harbor numerous aberrations that can act in additive, parallel, antagonistic, epistatic or synergistic fashion. Those interactions contribute to tumorigenesis, progression, metastasis, drug resistance or other life-threatening features. While these interactions can be weakly inferred from analysis of tumor sequence data, elucidating genetic interactions in vivo is essential for rapidly building a robust map of cancer development and to accelerate therapeutic developments. However, there are currently few effective tools for precise multigenic manipulation of cancer in vivo, limiting our scope for accurately dissecting these interactions. We endeavored to harness single-effector RNA-guided endonucleases (RGNs) for genome editing, parallel screening and in vivo modeling of human cancer. Recently, we generated a platform to systematically interrogate several hundred loci directly in vivo. To overcome current limitations in multigene editing and achieve more accurate control of simultaneity and sequentiality of multi-allelic tumor modeling, we utilized Cpf1, an RGN that can edit its target simply with crRNAs independent of tracrRNA thus allowing simultaneous editing of multiple genes with a single crRNA array. We developed a preliminary Cpf1-based crRNA array screening (CCAS) system in mammalian cells, and applied it in mouse models of progression and metastasis. In our first aim, we will perform validation and optimization of CCAS for in vivo double-knockout phenotyping of cancer co-drivers. We will establish its technical rigor, efficiency and specificity for simultaneous editing, as well as developing a set of computational pipelines for accurate calling of statistically significant gene pairs. We will apply this approach to study the genetic interactions of tumor suppressors found in lung cancer patients at Yale Cancer Center and Hospital, and identify potential co-drivers of metastasis to vital organs. In the second aim, we will carry out validation and optimization of a Cpf1-Flip system for sequential mutagenesis of cancer targets. We will demonstrate its broader applicability by testing clinically relevant gene sets identified from public studies of the genomics of metastasis as well as a large multi-sample metastasis dataset gathered on Yale cancer patients. We will then apply this methodology as an unbiased depletion screen to identify targets that are essential for survival in specific oncogenic backgrounds. We will develop novel versatile transgenic mouse strains and companion viral vectors for direct modeling of multigenic tumorigenesis in mice. We will combine these tools to enable high-throughput genetic interaction screening in healthy cells directly in the native organ to identify causative mutation pairs that drive tumorigenesis. We anticipate that developing and establishing these tools will transform multigenic tumor modeling and pre-clinical studies of human cancer, directly addressing NCI Provocative Question 4. These powerful toolkits will enable scientists to target any gene pairs or combinations simultaneously or sequentially, assessing the phenotypic outcome of their in vivo interactions in tumor progression, metastasis, synthetic lethality, drug sensitivity or other processes in cancer evolution.